Snowflake, a company specializing in the AI Data Cloud, has unveiled a new series of product innovations aimed at making it easier for businesses to develop data and artificial intelligence projects. These updates provide intuitive tools, a fully interoperable environment, and reliable AI agents that enable a faster journey from initial idea to production deployment.
Among the main enhancements is the general availability of Cortex Code, a native coding agent for data environments that automates and accelerates the entire enterprise development process. This agent is designed to understand and operate within the full context of corporate data. Also announced is the general availability of Semantic View Autopilot, an AI-driven service that simplifies the creation and governance of semantic views, giving AI agents a common interpretation of business metrics to ensure coherent and consistent results. Additionally, Snowflake has made advancements in Snowflake Postgres — with general availability coming soon — which will allow the world’s most widely used database to run natively in the AI Data Cloud, facilitating the unification of transactional, analytical, and AI workloads on a single secure platform.
These innovations were showcased at BUILD London, the company’s premier annual developer event, following the announcement of its multi-year partnership with OpenAI, valued at $200 million. This agreement reinforces their joint co-innovation and go-to-market strategy, integrating OpenAI’s models natively into Snowflake Cortex AI for Snowflake’s more than 12,600 global customers.
“For artificial intelligence to generate real value, it must move beyond experimental phases and be integrated into the systems teams use daily,” said Christian Kleinerman, Senior Vice President of Product at Snowflake. “With these new advancements, we are transforming how teams create and operate, embedding AI directly into the development lifecycle—preparing data for AI from the outset and helping organizations achieve tangible business impact. It’s a profound shift in how we work with data and AI, enabling reliable, governed, and scalable enterprise solutions.”
Snowflake introduces Cortex Code, an AI coding agent that understands enterprise data context to help teams build faster
With general availability of Cortex Code, users gain an agent that comprehensively understands and operates within the context of enterprise data. It empowers everyone—from data experts to domain specialists—to create data flows, analyses, and AI applications more quickly.
In its pursuit of making AI deliver real impact, organizational teams face mounting pressure to speed up without sacrificing reliability, accuracy, or scalability. To accelerate data and AI initiatives with greater confidence, teams need custom-designed tools that understand their data environments, simplify complex tasks, and enable sophisticated, reliable workflows via natural language.
Unlike generic coding assistants, Cortex Code understands the semantics of data, compute, governance, and user operations within Snowflake. It is customizable and interoperable, built to work wherever users operate—across Snowflake experiences and local development environments. It naturally integrates into existing workflows and supports the entire development lifecycle, from design and implementation to optimization and operations, all without compromising trust or security. Teams can use Cortex Code within the Snowflake platform through Cortex Code in Snowsight (coming soon), or within their preferred terminal or code editor, like VS Code or Cursor, via Cortex Code CLI (already generally available).
To further reduce friction slowing enterprise AI adoption and delivery, Snowflake has introduced new capabilities for more intuitive coding—driving how users build, deploy, and manage AI-driven data workflows across the tech stack, including a new integration with v0 of Vercel (coming soon). This allows employees—from developers to analysts—to create complete AI-powered applications using natural language, deployable securely within Snowflake via Snowpark Container Services.
Snowflake offers Semantic View Autopilot to ensure AI agents operate with shared business confidence definitions
Snowflake has introduced Semantic View Autopilot (now generally available), an AI-driven service that continuously learns from real user activity to ensure business logic is accurate and up-to-date. This helps minimize AI hallucinations while reducing the time to create semantic models from days to minutes, speeding time-to-market and providing a competitive edge. Built on Snowflake’s robust enterprise infrastructure, these innovations ensure that AI systems like Snowflake Intelligence are reliable, governed, and ready to operate securely at scale—working directly with organizations’ most valuable data.
Organizations often deploy AI agents in environments where business metrics are manually defined and governance is inconsistent, leading to systems lacking a shared business context. This fragmented and manual approach to building semantic layers creates a significant bottleneck for AI adoption, resulting in unreliable outcomes and eroding trust in AI systems.
Semantic View Autopilot addresses this challenge by automatically creating, optimizing, and maintaining governed semantic views, eliminating manual and error-prone semantic modeling. It builds upon Snowflake’s commitment to initiatives like the Open Semantic Interchange (OSI), which standardizes a shared semantic layer across the ecosystem. While OSI provides connectivity to share business logic across platforms, Semantic View Autopilot adds the intelligence to create and maintain it continuously—making it the connective layer for reliable, scalable AI across all data, wherever it resides.
Snowflake prepares enterprise data for AI with Snowflake Postgres and advanced innovations for open data interoperability
Snowflake has announced new improvements to Snowflake Postgres, now running natively in the AI Data Cloud, enabling users to easily access, ingest, and migrate data from anywhere to build what they need on a single secure platform.
Many organizations still keep their transactional and analytical databases isolated in separate systems—an old-fashioned approach that forces teams to rely on complex pipelines to connect these systems. This fragmentation incurs high costs, slows development, introduces risks, and delays access to critical insights.
Snowflake Postgres eliminates these pipelines by consolidating transactional, analytical, and AI use cases into a single enterprise-grade platform. Its full compatibility with open-source Postgres allows companies to migrate existing applications to Snowflake without code changes. Now, with Snowflake Postgres, teams can leverage AI and critical applications, analyze real-time performance and business trends using up-to-date operational data, and build AI-enabled features like recommendations or forecasts—all from the same data sources without moving data between systems.